Energy Optimization (EO) Pipeline

A MINLP optimization framework for energy system cost minimization, mirroring a production database workflow.

📁 Folder Structure

C:\Users\tnigam\Desktop\Python EO\
├── source/                          ⭐ Optimizer core
│   ├── Optimizer_MINLP.ipynb        (GEKKO MINLP solver)
│   ├── post_process_outputs.py      (DAG solver + CSV emitter)
│   ├── db_writer.py                 (Database output writer)
│   ├── feature_file_eo_v7_unified.xlsx  (Variables, Constraints, Formulas)
│   ├── Feature_File_main_v01.xlsx   (Legacy v6 reference)
│   └── tables_from_db/              (DB snapshots)
│
├── driver/                          ⭐ Scenario execution
│   ├── _run_driver.py               (CLI entry point)
│   ├── EO_All_Scenarios_Driver.ipynb (Notebook version)
│   └── README.md                    (Driver documentation)
│
├── outputs/                         📊 Results
│   ├── latest/                      (Most recent run results + scenario_runs/)
│   ├── archive/                     (Historical runs by timestamp)
│   └── logs/                        (Execution logs)
│
├── reports/                         📈 Report generation
│   ├── EO_Report_Generator.ipynb    (Generate PDF/Excel reports)
│   └── generated_reports/           (Output location)
│
├── app/                             🔧 Frontend/Dashboard
│   ├── eo-sensitivity-app/          (React sensitivity analysis app)
│   └── [coming: EO_FRONTEND_BACKEND integration]
│
├── docs/                            📚 Documentation
│   ├── DB_WRITEBACK.md              (Database export guide)
│   ├── PROJECT_STATE.md             (Handoff doc for cold start)
│   ├── EO_FRONTEND_BACKEND/         (Integration runbooks)
│   ├── db_pulls/                    (Database queries)
│   ├── excel_files/                 (Reference Excel files)
│   ├── PRODUCT CODING STANDARDS.pdf (Standards doc)
│   ├── .env & .env.example          (Config templates)
│   └── logs/                        (Execution history)
│
└── _obsolete/                       🗑️ Legacy files
    ├── bundle_backup/               (Old EO_AllScenarios_Bundle)
    ├── legacy_notebooks/            (Archived notebooks)
    └── build_scripts/               (Deprecated build utilities)

🚀 Quick Start

Run scenarios

cd driver/
python _run_driver.py baseline power_high fuel_low

Or edit scenarios in _run_driver.py and execute EO_All_Scenarios_Driver.ipynb

Output locations

• Per-scenario results: outputs/latest/scenario_runs/<scenario_name>/
• Cross-scenario summary: outputs/latest/eo_all_scenarios_TIMESTAMP.xlsx
• QC reports: outputs/latest/optimizer_qc_report_*.xlsx, python_vs_db_parity_*.xlsx

📋 Pipeline Flow

_run_driver.py 
  → reads scenario config
  → injects params into Optimizer_MINLP.ipynb
  → runs optimizer (IPOPT relax → round → APOPT lock)
  → calls post_process_outputs.py
  → writes 7 DB-schema CSVs + output_v3.xlsx
  → aggregates results to cross-scenario summary

🔧 Key Files

File	Role
source/Optimizer_MINLP.ipynb	3-stage GEKKO solver (never edit directly)
source/post_process_outputs.py	DAG post-optimizer, CSV/XLSX writer
source/feature_file_eo_v7_unified.xlsx	Variables (172), Constraints (44), Derived Eqs (117)
driver/_run_driver.py	Scenario parameterization & orchestration
reports/EO_Report_Generator.ipynb	Live optimizer + PDF report generation

📊 Outputs

• model_output.csv — Post-processed optimizer results (authoritative)
• output_v3.xlsx — Optimizer raw output with patched actuals/optimums
• 7 DB-schema CSVs — inferred_*, variables_*, constraint_*, etc.
• eo_all_scenarios_*.xlsx — Cross-scenario summary with all KPIs

🎯 Scenarios Available

Price variations:

• baseline — nominal rates
• power_low/high/vlow/vhigh — demand price sweep ($8–$100/MWh)
• power_40/50/55/60/100/200/350/400/500 — granular power rate sweep
• fuel_low/high/vlow/vhigh/10x — fuel cost variations
• dmw_low/high — demand margin variations

Infeasibility tests:

• infeas_3_boilers, infeas_all_turbines, infeas_1_boiler — constraint validation

Special:

• negative_power — test energy purchase scenario

📝 Notes

• First run may take 2–3 min per scenario (GEKKO compile + solve)
• Subsequent runs cache compiled models for ~30 sec per scenario
• Local optima observed at extreme power rates ($400+) — solution may be suboptimal
• All changes to optimizer logic go in source/post_process_outputs.py, never in copies

---

For detailed technical explanation of Variables, Constraints, and Derived Equations, see docs/PROJECT_STATE.md.